S
ugar sensing clearly plays an important role in plant cell metabolism and the expression of many genes, including several involved in photosynthesis and carbo-hydrate metabolism, is affected by sugars. Un-ravelling the molecular mechanisms of plant sugar sensing is of fundamental importance to understanding plant metabolism and might have biotechnological applications, for exam-ple, in manipulating partitioning of carbon between different storage products (starch, proteins, oils) in crop seeds and tubers. At pres-ent, however, our understanding of this pro-cess is far from complete, and it is worth reviewing the much greater (although incom-plete) body of work on sugar sensing in fungi.Sugar sensing and glucose repression in fungi
In Saccharomyces cerevisiae, genes required for growth on carbon sources other than glu-cose are repressed by the presence of gluglu-cose in the medium, and can be derepressed when it is removed. This is the phenomenon of glu-cose repression (also known as catabolite repression), which requires a mechanism for
sensing the availability of glucose1. In spite of
intensive study, the intracellular signals that mediate repression and derepression of genes in response to glucose in yeast remain un-known. Possible routes through which glu-cose could be sensed and a signal transduced are shown in Fig. 1. The signal for relief from catabolite repression in eubacteria is cyclic AMP, which also increases in response to glu-cose in S. cerevisiae and can regulate the expression of a few glucose-sensitive genes. However, a variety of evidence indicates that cyclic AMP is not a key component of the pathway controlling repression and derepres-sion of most glucose-regulated genes in this
organism1. The idea that glucose-6-phosphate
might be a signal appears to have originated with experiments that demonstrated that 2-deoxyglucose or glucosamine could mimic the
effect of glucose in gene repression2. Because
these hexose sugars were not metabolized at significant rates beyond the hexose phosphate stage, this suggested that the signal for glucose
repression was either the hexose itself or the hexose phosphate formed from it.
A flaw in this approach is that sugars that are rapidly phosphorylated by hexokinase, but for which the hexose phosphate is metabolized slowly or not at all, cause depletion of cellu-lar ATP. The hexose phosphate builds up to high levels in the cell, trapping phosphate and preventing rephosphorylation of ADP to ATP.
In our experience3, incubation of yeast with
the concentrations of 2-deoxyglucose used in
these studies2 results in a .95% decrease
in cellular ATP. We therefore regard ‘non-metabolizable’ sugars, such as 2-deoxyglu-cose, to be toxic compounds that reduce ATP to levels where normal cellular function might be seriously impaired. Interestingly, although
Holzer’s group2favoured the idea that changes
in hexose or hexose phosphate were the key signal, they also discussed changes in ATP levels as an alternative possibility.
The idea that the signal molecules were glucose and/or glucose-6-phosphate sug-gested that the enzyme hexokinase might be the sensor protein. The hexokinase hypothesis gained credence in the yeast-research com-munity when mutations (hex1) found to cause partially constitutive expression of glucose-repressed genes were mapped to the gene
HXK2, encoding the major hexokinase
iso-form PII (Ref. 4). Initial work with hxk2 alleles suggested that the glucose repression and hexo-kinase functions could be dissociated, and this led to the proposal that hexokinase PII had a glucose-sensing function independent of its
enzymic activity5. However, this did not stand
up to further investigation and there is, in fact, a good correlation between the overall hexo-kinase activity of different mutants and their
ability to exhibit glucose repression6,7. These
results suggest that hexokinase PII has a role in producing the signal molecule(s) but does not support the idea that it has a sensing role
per se. The glucose repression phenotype of hxk2 mutants might merely be because the
slow rate of glucose phosphorylation restricts overall glucose metabolism and ATP produc-tion when the major hexokinase isoform is missing or defective.
117
trends in plant science
perspectives
March 1999, Vol. 4, No. 3 1360 - 1385/99/$ – see front matter © 1999 Elsevier Science. All rights reserved. PII: S1360-1385(99)01377-1
Is hexokinase really a sugar
sensor in plants?
Nigel G. Halford, Patrick C. Purcell and D. Grahame Hardie
The molecular mechanisms by which plant cells sense sugar levels are not understood, but current models (adapted from models for sugar sensing in yeast) favour hexokinase as the primary sugar sensor. However, the hypothesis that yeast hexokinase has a signalling function has not been supported by more recent studies and the idea that hexokinase is involved in sugar sensing in plants has yet to be proven.
Fig. 1. Sugar sensing in a yeast cell. Although the intracellular signals involved in sugar sensing in yeast remain elusive, it is clear that hexoses, of which glucose is the most effec-tive, are perceived and that a functional sucrose nonfermenting-1 (SNF1) complex (con-taining the products of the SNF1 and SNF4 genes, and one of the products of the SIP1, SIP2 or GAL83 genes) is required.
Glucose
Glucose-6-phosphate
Ethanol Transporter
and/or receptor?
Glucose
Signal
SNF1 Complex
Downstream effects
Direct effects on metabolism Protein phosphorylation Indirect effects on metabolism
Regulation of gene expression
Hexokinase
Changes in AMP and/or ATP? Signal
Nucleus
Work in other fungi also suggests that hexo-kinase does not play a critical role in glucose repression. In Aspergillus nidulans, a mutant (frA1), which lacks hexokinase activity, al-though still expressing glucokinase, exhibits
normal glucose regulation of several genes8.
In Candida utilis, under conditions where hexo-kinase activity is reduced to 10% by the ad-dition of an inhibitor to the medium, glucose
repression of a-glucosidase is not affected9.
Although the nature of the signal molecules remains unclear, genes involved in reversal of glucose repression in S. cerevisiae are well characterized. In particular, a functional su-crose nonfermenting-1 (SNF1) gene is essen-tial for derepression of all glucose-repressed genes. The SNF1 gene encodes a protein-serine/
threonine kinase10, and the active kinase is a
high molecular mass complex (which we shall term the SNF1 complex) containing the prod-ucts of the SNF1 and SNF4 genes, and one of the products of the SIP1, SIP2 or GAL83
genes11. These are the homologues of the a, b
and gsubunits, respectively, of mammalian
AMP-activated protein kinase (AMPK)12. Both
SNF1 and SNF4 are essential for the reversal of
glucose repression. The products of the three separate genes, SIP1, SIP2 and GAL83,
repre-sent alternate forms of the bsubunit, giving
rise to at least three forms of the kinase com-plex. Surprisingly, all three can be deleted in
S. cerevisiae without apparent effect11, although
in the yeast, Kluyveromyces lactis, the FOG1
gene – encoding a bsubunit homologue – is
essential for reversal of glucose repression13.
Of several genes shown to act downstream from the SNF1 complex in S. cerevisiae, MIG1 encodes a DNA-binding protein that represses transcription of glucose-repressed genes in the presence of glucose. The amino acid sequence
of Mig1 contains several consensus sites14for
phosphorylation by the SNF1 complex, and mu-tation of these results in constitutive repression
of transcription15. Mig1 translocates from the
nucleus to the cytoplasm on glucose removal,
and this is associated with its phosphorylation16.
These results suggest that the SNF1 complex derepresses genes on removal of glucose (at least in part) by direct phosphorylation of Mig1, causing its translocation to the cytoplasm.
Regulation of AMP-activated protein kinase
Regulation of the mammalian AMPK com-plex is understood better than that of its yeast counterpart. AMPK is the downstream com-ponent of a kinase cascade, being activated by
an upstream kinase, AMPK kinase (AMPKK)17.
The system is activated in a highly sensitive
manner by elevation of 59-AMP and depletion
of ATP, via a complex mechanism involving activation of both AMPK and AMPKK, pro-motion of phosphorylation of AMPK and
in-hibition of its dephosphorylation12. The effects
of AMP are antagonized by high (mM) con-centrations of ATP, so that the system responds to changes in the AMP:ATP ratio rather than merely to changes in AMP. Because adenylate
kinase maintains its reaction (2ADP ↔ATP
1AMP) close to equilibrium, the AMP:ATP
ratio varies approximately as the square of the
ADP:ATP ratio18, and is a very sensitive
indi-cator of the cellular energy charge.
The AMPK cascade is activated by en-vironmental stresses that deplete cellular ATP
(Ref. 12) and also, at least in pancreatic b
cells, by glucose deprivation19. This raises the
intriguing possibility that the SNF1 complex might be activated in a similar manner in response to glucose deprivation of yeast, and that changes in AMP and ATP might be the signals responsible. The SNF1 complex was indeed found to be activated by glucose
re-moval3,20. It appears to be due to
phosphoryl-ation by an upstream kinase, and to be associated with large increases in the cellular
AMP:ATP ratio3. Although AMP does not
allosterically activate the SNF1 complex3,20, it
remains possible that it could activate the sys-tem through one of the other mechanisms established for the AMPK system, such as activation of the upstream kinase. Glucose-6-phosphate and a wide range of other sugar phosphates do not affect the activity of the
SNF1 kinase complex in vitro3.
Sucrose nonfermenting-1-related protein kinase 1s
Plant homologues of SNF1 and AMPK, SNF1-related protein kinase 1s (SnRK1), share con-siderable amino acid sequence identity with SNF1 (~67% in the protein kinase catalytic
domain)21. They complement snf1 mutants,
enabling them to use carbon sources such as glycerol, ethanol and sucrose, indicating that they can function within the glucose-sensing
signalling pathway in yeast22,23.
Is hexokinase a hexose sensor in plants?
The notion that hexokinase is a primary-sugar sensor and that it is capable of initiating a signal has been adopted enthusiastically by a substantial part of the plant science commu-nity24–26. The hypothesis has been tested in
118
trends in plant science
perspectives
March 1999, Vol. 4, No. 3
Fig. 2. Sugar sensing in a plant cell in a sink organ such as a seed or tuber. The predominant sugar transported to a plant sink cell is sucrose. It might be sensed itself, or it might be converted to hexoses either outside or inside the cell, which means that several different signalling pathways are possible. The components of these pathways have not been identi-fied, but an active sucrose nonfermenting-1-related protein kinase 1 (SnRK1) complex is required for sucrose-induced expression of the gene encoding sucrose synthase in potato, suggesting that SnRK1 performs a role analogous to that of sucrose nonfermenting-1 (SNF1) in yeast.
Sucrose
Hexose phosphates Transporter
and/or receptor?
Sucrose
Signal Signal Receptor Hexoses
SnRK1 Complex
Downstream effects
Direct effects on metabolism
Protein phosphorylation Indirect effects on metabolismRegulation of gene expression
Signal Hexokinase
Changes in AMP and/or ATP?
Nucleus
Cytoplasm Cell Wall
Transporter and/or receptor?
Hexoses
Hexoses
transgenic Arabidopsis by antisense and over-expression of hexokinase. Plants with higher hexokinase levels were more sensitive to glucose, using photosynthetic gene expression as a marker for glucose repression, whereas plants with lower activity were less
sensi-tive27. However, these results could be
ex-plained in a similar way to those involving yeast hexokinase mutants; that is, alterations in hexokinase activity might simply affect the rate of metabolism of hexoses and there-fore the production of ATP. In this latter model, hexokinase is a rate-limiting enzyme for glucose metabolism, but not a sugar sen-sor per se.
Another line of evidence that supports the hypothesis that hexokinase is a sugar sensor involved treating maize protoplasts with dif-ferent sugars, again using photosynthetic gene
expression as a marker for sugar sensing28.
Similar experiments were performed on a cucumber cell line using isocitrate lyase- and
malate synthase-gene expression as markers29
(these two genes are repressed by sugars in both plants and yeast). In the maize protoplasts and the cucumber callus cultures 2-deoxyglu-cose causes a marked degree of gene repres-sion, whereas the non-phosphorylatable sugar, 3-O-methyl glucose does not. These results were interpreted as evidence that metabolism beyond the hexose-6-phosphate is not required
to generate the signal28,29. However, the same
caveat applies to the use of ‘non-metabolizable’ sugars in plants as in yeast: even if they pro-duce only subtle alterations in cellular ATP (and hence ADP and AMP), they could, as a secondary consequence, affect the level of al-most any cellular metabolite.
It has also been reported that the introduc-tion of glucose into cells via electroporaintroduc-tion causes repression, whereas introduction of
glucose-6-phosphate does not28. This could be
taken as evidence that phosphorylation of glu-cose is not necessary to generate the signal. However, in our view, this is not a valid com-parison because glucose would still be able to enter the cells by facilitated transport once the electroporation pores had sealed, whereas glu-cose-6-phosphate would only enter for the brief period when the pores were open.
Another reason for caution in interpreting these results is that they were performed on protoplast cultures or cell lines. We have measured very high levels of SnRK1 activity in plant material grown in culture. For exam-ple, activity in potato mini-tubers and callus cultures is ~50- and 25-fold higher, respec-tively, than in mature leaves, and eight- and four-fold higher, respectively, than in any
tis-sue from whole potato plants30. This elevated
SnRK1 activity is likely to have a significant effect on the signalling pathway.
Doubts on a role for hexokinase in sugar sensing also come from evidence that high
hexose levels in the presence of hexokinase do not necessarily generate a signal. Ex-pression of yeast invertase in the cytosol of transgenic tobacco plants results in high lev-els of glucose and fructose but, in spite of the fact that hexokinase is a cytosolic en-zyme, the elevated levels of hexoses do not
appear to be sensed31,32. However, when the
invertase is targeted to the vacuole or apoplast, the high levels of glucose and fructose that are produced in these compartments are sensed and changes in gene expression are observed. These results suggest that the signal is sensed at a membrane location rather than in the cytoplasm.
Sucrose sensing in plants
One clear difference between yeast and plant cells is the importance of sucrose as a carbon source. Whereas for yeast, sucrose is merely one carbon source among many that can be turned to if insufficient glucose is available, for plants, sucrose is the major transported sugar, linking carbon source or-gans (mature leaves) with carbon sinks (de-veloping leaves, roots and storage organs such as seeds and tubers; Fig. 2). Of course, sucrose and hexose levels might be linked, depending on the sucrose-metabolizing en-zymes present in the cell, and sucrose and glucose treatments have had similar effects in
some experiments. For example, b-amylase
gene expression can be induced by sucrose
or glucose33. However, it is clear that sucrose
itself can act as an effector in plant cells, generating a signal independent of that gen-erated by glucose. Sucrose synthase gene expression in potato cell cultures, for ex-ample, is induced by sucrose but not by glu-cose34,35. This induction appears to require
SnRK1 activity, because antisense inhibition of SnRK1 gene expression in potato results in a dramatic decrease in sucrose synthase gene expression in tubers and in excised leaves
incubated with sucrose36. It seems likely that
SnRK1 regulates the expression of other genes as well, in a role analogous to that of SNF1 in yeast, and we have given SnRK1 a central position in Fig. 2. However, this is not proven.
Further evidence that sucrose and glucose can affect sugar sensing in different ways has been obtained from experiments on
de-veloping cotyledons of Vicia faba37. During
the early stages of cotyledon development, a seed coat-associated invertase cleaves su-crose to maintain high levels of hexoses in the cotyledon. This activity disappears as the cotyledon cells begin to differentiate into storage tissue, and the cellular hexose levels fall and sucrose levels rise. This coincides with a steep rise in sucrose synthase activity and the onset of starch biosynthesis in the cotyledons.
Concluding remarks
It is now clear that there are multiple sugar-sensing pathways in plants and that sucrose and hexoses can initiate different signals in some tissues in a way that has not been shown in yeast. The term ‘sugar sensing’ is, therefore, imprecise as far as plants are concerned and terms such as hexose and sucrose sensing are more appropriate. Clearly, some of these path-ways do not involve hexokinase and the hypothesis that hexokinase plays any role in hexose sensing in plants has yet to be sup-ported by unequivocal experimental evidence.
Acknowledgements
IACR receives grant-aided support from the UK Biotechnology and Biological Sciences Research Council (BBSRC). D.G.H. was sup-ported by a Project Grant from the BBSRC.
References
1 Gancedo, J.M. (1998) Yeast carbon catabolite
repression, Microbiol. Mol. Biol. Rev. 62, 334–361
2 Witt, I., Kronau, R. and Holzer, H. (1966)
Repression von Alkoholdehydrogenase, Maltatdehydrogenase Isocitratalyase und Malatsynthase in Hefe durch Glucose, Biochim.
Biophys. Acta 118, 522–537
3 Wilson, W.A., Hawley, S.A. and Hardie, D.G.
(1996) The mechanism of glucose
repression/derepression in yeast: SNF1 protein kinase is activated by phosphorylation under derepressing conditions, and this correlates with a high AMP:ATP ratio, Curr. Biol. 6, 1426–1434
4 Entian, K.D. (1980) Genetic and biochemical
evidence for hexokinase PII as a key enzyme involved in carbon catabolite repression in yeast,
Mol. Gen. Genet. 178, 633–637 5 Entian, K.D. and Frohlich, K.U. (1984)
Saccharomyces cerevisiae mutants provide
evidence of hexokinase PII as a bifunctional enzyme with catalytic and regulatory domains for triggering carbon catabolite repression,
J. Bacteriol. 158, 29–35
6 Ma, H. et al. (1989) The residual enzymatic
phosphorylation activity of hexokinase II mutants is correlated with glucose repression in
Saccharomyces cerevisiae, Mol. Cell. Biol. 9,
5643–5649
119
trends in plant science
perspectives
March 1999, Vol. 4, No. 3
Nigel G. Halford*and Patrick C. Purcell are at IACR-Long Ashton Research Station, Dept of Agricultural Sciences, University of Bristol, Long Ashton, Bristol, UK BS41 9AF. D. Grahame Hardie is at the Dept of Biochemistry, Dundee University, MSI/TWB Complex, Dow Street, Dundee, UK DD1 4HN.
*Author for correspondence
7 Rose, M., Albig, W. and Entian, K.D. (1991)
Glucose repression in Saccharomyces cerevisiae is directly associated with hexose
phosphorylation by hexokinase PI and PII, Eur. J.
Biochem. 199, 511–518
8 Ruijter, G.J. et al. (1996) Characterization of the Aspergillus nidulans frA1 mutant: hexose
phosphorylation and apparent lack of
involvement of hexokinase in glucose repression,
FEMS Microbiol. Lett. 139, 223–228 9 Espinel, A.E., Gómez-Toribio, V. and
Peinado, J.M. (1996) The inactivation of hexokinase activity does not prevent glucose repression in Candida utilis, FEMS Microbiol.
Lett. 135, 327–332
10 Celenza, J.L. and Carlson, M. (1986) A yeast
gene that is essential for release from glucose repression encodes a protein kinase, Science 233, 1175–1180
11 Jiang, R. and Carlson, M. (1996) The Snf1
protein kinase and its activating subunit, Snf4, interact with distinct domains of the Sip1/Sip2/Gal83 component in the kinase complex, Mol. Cell. Biol. 17, 2099–2106
12 Hardie, D.G., Carling, D. and Carlson, M.
(1998) The AMP-activated/SNF1 protein kinase subfamily: metabolic sensors of the eukaryotic cell? Annu. Rev. Biochem. 67, 821–855
13 Goffrini, P. et al. (1996) FOG1 and FOG2 genes,
required for the transcriptional activation of glucose-repressible genes of Kluyveromyces
lactis are homologous to GAL83 and SNF1 of Saccharomyces cerevisiae, Curr. Genet. 29,
316–326
14 Dale, S. et al. (1995) Similar substrate
recognition motifs for mammalian AMP-activated protein kinase, higher plant HMG-CoA reductase kinase-A, yeast SNF1, and mammalian calmodulin-dependent protein kinase I, FEBS
Lett. 361, 191–195
15 Östling, J. and Ronne, H. (1998) Negative control
of the Mig1p repressor by Snf1p-dependent phosphorylation in the absence of glucose, Eur. J.
Biochem. 252, 162–168
16 DeVit, M.J., Waddle, J.A. and Johnston, M.
(1997) Regulated nuclear translocation of the Mig1 repressor, Mol. Biol. Cell 8, 1603–1618
17 Hawley, S.A. et al. (1996) Characterization of the
AMP-activated protein kinase kinase from rat liver, and identification of threonine-172 as the major site at which it phosphorylates and activates AMP-activated protein kinase, J. Biol.
Chem. 271, 27879–27887
18 Hardie, D.G. and Carling, D. (1997) The
AMP-activated protein kinase: fuel gauge of the mammalian cell? Eur. J. Biochem. 246, 259–273
19 Salt, I.P. et al. (1998) AMP-activated protein
kinase is activated by low glucose in cell lines
derived from pancreatic bcells, and may regulate
insulin release, Biochem J. 335, 533–539
20 Woods, A. et al. (1994) Yeast SNF1 is functionally
related to mammalian AMP-activated protein kinase and regulates acetyl-CoA carboxylase in
vivo, J. Biol. Chem. 269, 19509–19515 21 Halford, N.G. and Hardie, D.G. (1998)
SNF1-related protein kinases: global regulators of carbon metabolism in plants? Plant Mol. Biol. 37, 735–748
22 Alderson, A. et al. (1991) Complementation of snf1, a mutation affecting global regulation of
carbon metabolism in yeast, by a plant protein kinase cDNA, Proc. Natl. Acad. Sci. U. S. A. 88, 8602–8605
23 Muranaka, T., Banno, H. and Machida, Y. (1994)
Characterization of the tobacco protein kinase NPK5, a homologue of Saccharomyces
cerevisiae SNF1 that constitutively activates
expression of the glucose-repressible SUC2 gene for a secreted invertase of S. cerevisiae, Mol.
Cell. Biol. 14, 2958–2965
24 Jang, J.C. and Sheen, J. (1997) Sugar sensing in
higher plants, Trends Plant Sci. 2, 208–214
25 Taylor, C.B. (1997) Sweet sensations, Plant Cell
9, 5–19
26 Smeekens, S. (1998) Sugar regulation of gene
expression in plants, Curr. Opin. Plant Biol. 1, 230–234
27 Jang, J.C. et al. (1997) Hexokinase as a sugar
sensor in higher plants, Plant Cell 9, 5–19
28 Jang, J.C. and Sheen, J. (1994) Sugar sensing in
higher plants, Plant Cell 6, 1665–1679
29 Graham, I.A., Denby, K.J. and Leaver, C.J.
(1994) Carbon catabolite repression regulates glyoxylate cycle gene expression in cucumber,
Plant Cell 6, 761–772
30 Man, A.L. et al. (1997) Potato SNF1-related
protein kinase: molecular cloning, expression analysis and peptide kinase activity measurements, Plant Mol. Biol. 34, 31–43
31 Herbers, K. et al. (1996) Systemic acquired
resistance mediated by the ectopic expression of invertase: possible hexose sensing in the secretory pathway, Plant Cell 8, 793–803
32 Heineke, D. et al. (1994) Accumulation of
hexoses in leaf vacuoles: studies with transgenic tobacco plants expressing yeast-derived invertase in the cytosol, vacuole or cytoplasm, Planta 194, 29–33
33 Mita, S., Suzuki-Fujii, K. and Nakamura, K.
(1995) Sugar-inducible expression of a gene for
b-amylase in Arabidopsis thaliana, Plant
Physiol. 107, 895–904
34 Salanoubat, M. and Belliard, G. (1989) The
steady-state level of potato sucrose synthase mRNA is dependent on wounding, anaerobiosis and sucrose concentration, Gene 84, 181–185
35 Fu, H. and Park, W.D. (1995) Sink- and
vascular-associated sucrose synthase functions are encoded by different gene classes in potato, Plant
Cell 7, 1369–1385
36 Purcell, P.C., Smith, A.M. and Halford, N.G.
(1998) Antisense expression of a sucrose non-fermenting-1-related protein kinase sequence in potato results in decreased expression of sucrose synthase in tubers and loss of sucrose-inducibility of sucrose synthase transcripts in leaves, Plant J. 14, 195–202
37 Weber, H. et al. (1996) Sucrose metabolism
during cotyledon development of Vicia faba L. is controlled by the concerted action of both sucrose phosphate synthase and sucrose synthase: expression patterns, metabolic regulation and implications for seed development, Plant J. 9, 841–850
120
trends in plant science
perspectives
March 1999, Vol. 4, No. 3
Agree or disagree?
Perspectives articles are your opportunity to express personal views on topics of interest to plant scientists. The range of appropriate subject matter covers the breadth of plant biology: historical, philosophical or political aspects; potential links between hitherto disparate fields; or simply a new view on a subject of current importance. Specific issues might include: the allocation of funding; commercial exploitation of data; the gene sequencing projects; genetic engineering; the impact of the internet; education issues; or public perception of science. Current developments in other disciplines may also be relevant.
If you wish to write for the perspectives section, please contact the Editor before embarking on your manuscript, with a 200-word proposal and a list of key recent references
Plants, unlike animals, continue to pro-duce novel organs and tissues during their post-embryonic phase1,2. This pro-longed morphogenetic activity could therefore be characterized as ‘continued embryo-genesis’3. The dormant state of the embryo in the mature seed does not mark the end of em-bryo development, but is part of the develop-mental continuum and assists seed dispersal4
. Although all plant development could be re-garded as a continuum, at some point the apical meristem (AM) starts to function as a highly autonomous unit1–3,5
. From that moment on-wards the AM can be viewed as a self-regulating structure, the activity of which gives rise to the shoot1,2,5. However, this view is not universally shared, and the AM is also regarded as a group of undifferentiated and dividing cells that are patterned by information sent out from the surrounding differentiated cells6. The AM may thus be viewed as either active (autonomous) or passive (non-autonomous) in pattern formation and morphogenesis.
Autonomous role of the apical meristem
The most convincing arguments for AM autonomy come from microsurgery experi-ments1–3. When the entire apex is excised and the explant grown in vitro, the AM simply con-tinues its morphogenetic activity. It appears that the highly specialized organization of the AM makes this possible. For example, when parts of the AM are surgically excised, the remaining cells reorganize themselves and re-establish a functional AM, even when only a small group of cells remain1. Similarly, sep-aration of the AM into two, by a vertical in-cision, results in the reorganization of the halves and the formation of two new meri-stems1–3. In each half a new centre is formed at the AM periphery that is distinct from the outline of the original centre1. This behaviour reveals some crucial aspects of AM organi-zation; it suggests – in terms of the positional information model7– a reprogramming of positional values or cellular ‘identities’ (mor-phallaxis) rather than the reconstruction of missing values by extrapolation (epimorpho-sis). In other words, the identities of the AM cells are not fixed, but can be reset by changes in the communication network established
with the neighbouring cells. The stable orga-nization of ‘the AM as a whole’1must there-fore be based on the stable organization of the communication network. In general, it is a characteristic of model networks with a feed-back architecture that the isolated parts con-serve the structure of the whole; the network underlying AM integrity could therefore, have such a structure. When the information about ‘the AM as a whole’ is lost, the individ-ual cells might reiterate their developmental potential and use an indirect route for meri-stem formation via callus and somatic em-bryogenesis. The latter occurs naturally at the base of apical explants1, where many cells are released from tissue constraints. If the AM were to behave as a passive group of cells6, this emergent property of the isolated dome would be hard to explain. Instead, the behav-iour of the AM might best be explained on the basis of self-organization of the supracellular network5,8.
Frameworks for networks
The communication network in the AM is sus-tained in a proliferating cellular context, which nonetheless shows regular and robust patterns. These cellular patterns are meaningful because they convey information about the underlying supracellular network. Early studies attempted to understand the AM in terms of cell division patterns. A useful framework for such studies
was the tunica–corpus concept9, which defines the tunica as the stable outer cell layer. How-ever, it appeared that meristems are variable in this respect and could either lack the entire tunica layer, possess more than one layer, or have an unstable layer. Gymnosperms appeared to lack a real tunica, but instead had a clear radial organization with distinct cytohistological dif-ferences between central and peripheral AM cells. This ‘cytohistological zonation’ pattern was subsequently found in many angiosperms, and was thought to be a better model for the AM than the tunica–corpus concept1. Cytohisto-logical zonation is still regarded as the only or-ganizational pattern that has a clear significance for the understanding of AM function1,3,4,10. The seemingly facultative nature of the tunica suggests that its presence in angio-sperms is more a peculiarity than an organi-zational feature1,3,10. The positional specification of cells in the angiosperm AM has been re-garded as further confirmation that the tunica might not play a role in morphogenesis10
. This view appears to be reinforced by the obser-vation that ferns and lower vascular plants, which entirely lack a tunica or superficial layer, are perfectly capable of forming a shoot10. However, the fact that a stable tunica is re-stricted to angiosperms9, whereas ferns do not possess even a superficial layer1, can be viewed from a perspective that assigns significance to the tunica. Because major plant taxa have very different shoot designs, it seems plausible that shoot construction is somehow dependent on the organization of the AM (i.e. on the particu-lar network of celluparticu-lar connections11
). This net-work of cell connections is determined by the number and position of initial cells and by the pattern in which they divide. These aspects ap-pear to be different in the major plant taxa, and three main types of AM have been discerned: the monoplex, simplex and duplex, which are characteristic of ferns and lower plants, gymno-sperms, and angiogymno-sperms, respectively9
(Fig. 1).
31
trends in plant science
perspectives
January 1999, Vol. 4, No. 1 1360 - 1385/99/$ – see front matter © 1999 Elsevier Science. All rights reserved. PII: S1360-1385(98)01362-4
Networks for shoot design
Chris van der Schoot and Päivi Rinne
The intrinsic capacity of the shoot apical meristem for self-regulation and the positional specification of its cells implies the existence of an elaborate and versatile communication network. We propose a model that pictures this network as a system of overlapping signal circuits, which support local tasks as well as coordinating indeterminate shoot development.
Fig. 1. Types of shoot apical meristem (AM): (a) the monoplex; (b) the simplex; (c) the duplex
which are characteristic of the majority of ferns and lower vascular plants, gymnosperms and angiosperms, respectively. These distinct types of AM organization arise due to differences in the number and position of initial cells, and the directions in which they divide. (a) The monoplex AM has a single top cell, which is often tetrahedral and produces daughter cells by lateral division. (b) The simplex AM has a zone of initial cells in an unstable superficial layer, where cells can divide in both directions. (c) The duplex AM has two distinct zones of ini-tial cells which give rise to two lineage compartments, tunica and corpus. As a result, the duplex AM has two separate cell production zones with distinct cellular features and growth dynamics. The cells of the tunica, as a rule, divide exclusively anticlinally. Modified after Newman; See Ref. 9.
